Multi-Core CPU Power Control via Usage Estimation and Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-core systems, existing power control methods lead to unnecessary power consumption by supplying minimum power to idle CPU cores and increase calculation complexity due to the need to calculate all multiple power policies.
Innovation Solution
A method and apparatus for efficient power control in multi-core systems that adaptively apply frequency/voltage control and core on/off schemes, estimating CPU usage for the next time interval, calibrating based on system information, and selecting a power control mode to minimize overhead and optimize power policy selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If minimum power is supplied to a core in a standby state, then the core remains ready for immediate operation, but unnecessary power consumption occurs
Solution Approach 1:
The patent implements dynamic power management by transitioning CPU cores between different power states (C0, C1, C2, C3) based on real-time workload analysis. The system dynamically adjusts power supply to cores, switching from full power (C0) to reduced power (C1-C3) states when cores are in standby, thereby reducing unnecessary power consumption while maintaining the ability to quickly resume operation when needed.
2Adaptability or versatility
If all multiple power policies are calculated to determine a power policy, then the most suitable power policy can be selected, but the complexity of calculation increases
Solution Approach 1:
The patent extracts and analyzes only the critical factors from the complete set of power policy parameters. Instead of calculating all possible power policies, the system identifies key workload characteristics (CPU usage patterns, prediction accuracy metrics) and uses these extracted features to determine the most suitable power policy, thereby reducing calculation complexity while maintaining adaptability.
Solution Approach 2:
The patent applies partial action by implementing a tiered approach to power policy calculation. The system performs basic workload analysis for all cores, then applies more sophisticated prediction algorithms only where needed based on initial assessment results. This selective application of calculation intensity reduces overall computational complexity while still achieving accurate power policy selection.
3Productivity
If the operation frequency of a CPU is increased to improve performance, then processing speed improves, but power consumption increases
Solution Approach 1:
The patent implements dynamic frequency and voltage adjustment based on real-time CPU workload analysis. The system continuously monitors CPU usage patterns and adjusts the operation frequency and voltage of CPU cores dynamically - increasing frequency when workload demands high processing speed, and reducing frequency when workload is light, thereby optimizing the balance between productivity and power consumption.
Data Source
AI summary
A power control method of a Central Processing Unit (CPU) in a multi-core system. The power control method includes acquiring current usage information of the CPU and system information, estimating a CPU usage of a next time interval based on the acquired current usage information, calibrating the estimated CPU usage of the next time interval based on the acquired system information, and determining a power control mode based on at least one of the acquired system information and the calibrated CPU usage of the next time interval.